Cryptocat, Now on iPhone | Hacker News Reader

Cryptocat, Now on iPhone

blog.crypto.cat

64 pointse1ven34 comments

Help me understand why I should rely on the privacy of this app, if the OS on which it's running is untrusted. Wouldn't it be trivial for Apple (or iOS malware) to record keystrokes?

Short of a serious undisclosed bug in iOS, malware should not be able to record keystrokes from other apps.

But you are right in that if you don't trust the OS manufacturer not to be snooping on you, you can't trust the app. In addition to logging keystrokes they could be logging the contents of UI text controls as they are set. Even if you implement your own custom keyboard and UI controls and your own encryption prior to sending anything over a socket they could still be snapshotting the phone's framebuffer regularly. There's no way to reliably hide data that you can see visually (or data that has to be unencrypted in memory at some point, even if you don't see it) from the lowest levels of the OS.

Having said all of that, they probably aren't doing any of this, but they could.

From a week ago, a proof of concept of this kind of bug, "Background Monitoring on Non-Jailbroken iOS 7 Devices": http://www.fireeye.com/blog/technical/2014/02/background-mon...

> iOS7 provides settings for "background app refresh". Disabling unnecessary app's background refreshing contributes to preventing the potential background monitoring. However, it can be bypassed. For example, an app can play music in the background without turning on its "background app refresh" switch. Thus a malicious app can disguise itself as a music app to conduct background monitoring.

I doubt this would get into the App store, but that's not the point, is it?

Of course, the idea is that the app would actually do something useful, would do that useful thing correctly, and only activate the malicious code for targetted users (none of which would be Apple testers).

It's an improvement over using a less secure chat program on the same OS.

Also, even if you use your own homemade ultra secure OS, you will still be running it on top of the closed source & exploitable broadband chip.

Not necessarily; some APs and BPs communicate over USB rather than the BP acting as a hypervisor.

If you think apple is spying on you through your phone, then change phone. If you don't, then it's a strong improvement over non-encrypted chats. If you are doubtful, then I would still get it because it's an improvement and thus why not, just in case Apple is not spying.

You could say the same about any current mobile OS. I think that having this kind of apps is better than not having them at all. Nothing in this life is 100% secure.

Downloaded it, but have no one to chat with.

I created a room called: hackernews

Its group fingerprint is: BBD398AA E131E1644 6EF77D4E A2CDC074 F497ED37

I'm not sure the group fingerprint is required. I've noticed if you logout and login with the same room name, it generates a different fingerprint.

I joined, but the fingerprint is different... :/

A unique fingerprint is issued to each individual.

Fingerprints are for verifying your own identity in the group chat context. They confirm your own identity to others.

It's hard for new users to understand this — the blog post mentions that Cryptocat is trying to make cryptographic authentication less confusing.

b here, well worth a look, buddies list seems to be crashy, but very well designed with minimal features. worth checking out.

magikarp, you're hellbanned. Apparently after your last comment about HN censorship, so... a little irony there I guess.

I went through this while experimenting with cryptocat, it's confusing as hell. the group fingerprint is your fingerprint, in that chatroom, for that session.

I peeked at some of the issues on Github and it looks like they're adding challenge-response for OTP verification so you can have a way to prove identity without an active out of band channel.

Alternately: https://chatsecure.org

I can never remember, is this the one that tptacek likes, or is that the other one?

You're thinking of TextSecure by Open WhisperSystems.

I like what Whisper is doing, and deeply respect the crypto talent that went into building it. Personally, I think the soundness of the cryptography is the most important attribute of a cryptographically secure chat application. Other people think other things.

Thinking correct things is more valuable than thinking popular things. They'll come around to your point of view in time.

"The market can stay irrational longer than you can stay solvent"

How is that relevant in this discussion?

Quoting the post I was replying to:

> They'll come around to your point of view in time.

The quote from Keynes was merely a slightly whimsical way to make the point that people can hold false beliefs for a very long time, and that it's probably better not to personally invest in the idea that they will come around by themselves in any specific timeframe.

No, this is the one written by a guy that just wanted to make "secure chat" accessible. And went on to have plenty of flaws. (And no, Telegram is another one, where they came up with their own scheme and brushed off all criticism by saying they had some math PhDs so that's that.) IIRC the web version does JS crypto and tries to sell that as more secure.

Cryptocat and Telegram seem to be very popular though, proving again that technical superiority isn't really a good indicator of product success. (WhatsApp had terrible security too, at least for a while, but they never went on about how secure they were.)

I think that's unfair - both text secure and cryptocat are interested in making secure chat more accessible.

Making security applications accessible is an important problem to try and solve. Security is difficult and I think they were fairly up front with the potential issues with what they were building.

They've also been doing it for a few years now, the code is open source, they've had professional security audits - what more could they possibly be doing?

Your parent was talking about Telegram, not TextSecure. The quality of the formers crypto is much more controversial than the latter.

TextSecure is solid as far as I know. What I meant with my comment is that Cryptocat/Telegram put marketing and accessibility over security. They rather sell a veneer of security in a cute package instead of actual security, along with the difficult problems it brings. Both products ignored (and actually told off) expert advice, which is quite telling.

IIRC, TextSecure was created by crypto expert[s].

Super simple interface, I like it. However it seems to be lacking an obvious way to leave your current conversation and enter a new one.

Not very secure if anyone can just waltz into your chat room.

Are you publicly posting your chat room's (not simple to guess) name?

It also supports direct messaging apparently. Click on buddies and then the user.

For everyone looking for a chatroom try "lobby".

Why does anyone still care about Cryptocat?

Reply on news.ycombinator.com